C39 - Multiple or Simultaneous Equation Models; Multiple Variables: OtherReturn

We estimate that the Unobserved Economy (UnEc) share in Portuguese GDP rose to a maximum of 34.37% in 2022. The results suggest that cutting the tax burden and improving the efficacity of social benefits could reduce UnEc and foster inclusive growth. The rising UnEc share MIMIC estimates found in our single-country study, with variables suited to Portugal, contrast with the downward trend (from 22.2% in 2003 to 16.5% in 2021) in the MIMIC multi-country study by Schneider (2021), who uses different variables - mostly reflecting the economic structure of advanced economies - and does not report if and how country heterogeneity is dealt with. Compared to other recent single-country estimates for the European Union periphery, Schneider (2021) also underreports the UnEc share of (at least) Greece, Poland and Lithuania. Thus, we recommend that authorities prefer UnEc single-country estimates in their analysis and policy design and be aware of mentioned issues in multi-country results.

The aim of the article is to introduce a complex econometric model of cash-lows for the Czech life insurance market. Namely, technical-actuarial links among insurance variables observed in annually published summary balance sheets of life insurers are described by means of an econometric system of linear simultaneous equations. The suggested model is statistically veri ed and thus it can provide useful economic interpretations. Further, adjusted residual bootstrapping is introduced in this context as a straightforward alternative which can solve possible problems with questionable asymptotic distribution properties of residuals. This technique can be applied e.g. for signi cance testing purposes. Finally, an important practical illustration of scenario analysis is considered. Such an analysis might be really useful, e.g. for internal calculations of the Czech life insurers, nancial planning or stress testing in the framework of Solvency II. Two general approaches are presented: deterministic and stochastic. The second one is capable of delivering various empirical probabilities concerning possible future developments.